What are approximate numeric data types supported in SQL Server Transact-SQL?

Approximate numeric data types are used to hold numeric values with floating scales.

There are 3 different approximate numeric data types supported in SQL Server Transact-SQL:

1. FLOAT - Used to hold values with different precisions and floating scales specified in the format of FLOAT(p), where p is the precision presenting the number of bits used to store the mantissa part. For example, FLOAT(12) stores floating numbers with a 12-bit mantissa.

2. DOUBLE PRECISION - Use to hold values with a large precision and floating scales using 8-byte storage to store floating numbers in the ranges of -1.79E+308 to -2.23E-308, 0 and 2.23E-308 to 1.79E+308. DOUBLE PRECISION is identical to FLOAT(53).

3. REAL - Use to hold values with a normal precision and floating scales using 4-byte storages to store floating numbers in the ranges of -3.40E+38 to -1.18E-38, 0 and 1.18E-38 to 3.40E+38.

Here are some good examples of approximate numeric values:

PRINT 9.22337203685e+010; -- FLOAT(53)
PRINT 9.22337e+010; -- FLOAT(24)

⇒ Precision and Rounding of FLOAT Values in SQL Server Transact-SQL

⇐ Overflow and Rounding on NUMERIC Values in SQL Server Transact-SQL

⇑ Variables and Data Types in SQL Server Transact-SQL

⇑⇑ SQL Server Transact-SQL Tutorials